Person. Q lnd. Oifl Vol. I, pp.

317 to 333
0 Pergamon Press Ltd 1980. Printed in Great Britain

PERSONALITY
FACTORS IN THE EYSENCK
PERSONALITY
QUESTIONNAIRE
PAUL BARRETTand PAUL KLINE
University of Exeter, Devon, England
(Received 19 March 1980)

Summary-Hierarchical
factor analyses were carried out on raw scored EPQ data yielded from
two samples of subjects, a Gallup adult quota sample of 1198 subjects and 406 university undergraduates. Each sample was split by sex providing six data sets for analysis. Results at the second
order level clearly indicated the recovery of practically all the E, N and L items as three distinct
factors, P item recovery differed among males and females. Adult female and student female
subject groups did not show clear recovery of a substantial majority of P items. Although P, as a
factor, did not appear recognisably in the male student sample, this was interpreted as being due
either to the low subject:variable ratio and/or to the low subject number (N = 171) and its
resultant lack of definition of initial P item variance. The P factor appeared most strongly as a
second order in the total Gallup, total student, and male Gallup sample.

The Eysenck Personality Questionnaire (EPQ) is a relatively new personality inventory,

introduced by Eysenck and Eysenck (1975) as a measure of Extraversion
(E),
Neuroticism/Anxiety (N), Psychoticism (P) and including a Social Desirability scale (L)
(Eysenck and Eysenck, 1976; Buros, 1978). The location of factors in hierarchical factor
analysis has, however, been open to question, Eysenck (1978) claiming that they generally
appear at the 2nd or 3rd order, although they may also appear at the 1st order (Eysenck
and Eysenck, 1968).
Recently Loo (1979) presented results from a hierarchical item factoring of the EPQ,
using 262 university students as subjects. He extracted sixteen oblique 1st order factors,
six 2nd order, and two 3rd order factors; principal components being the extraction
method employed at all levels. Only at the 2nd order did he find a factor akin to L, E, N
or P, this factor he named Social Extraversion. However, there are reasons to suggest
that the results are misleading. On a general level, there was a lack of information as to
the criteria adopted in the factoring procedures, in addition there was careless naming of
factors at the 1st and 2nd order levels, e.g. 1st order Impulsivity, Cooperative Sociability;
2nd order Psychopathy. Specifically, no information was provided as to the tests used for
ceasing to extract factors from any of the orders; grounded (Hendrickson and White,
1966) or non-grounded (Cattell and White, 1962) higher order rotation procedures were
not specified; apparently no attempt was made to seek a maximum hyperplane count
(HC) for each oblique rotation, although the HCs reported were so high that this is
probably not so important; and finally no mention was made of oblique factor intercorrelations at any level. Loo also used a strange procedure for his higher order factor
interpretation, he interpreted these factors on the basis of Varimax solutions, then,
assuming that they would remain the same, obliquely rotated them to obtain correlations
for further factoring; no test or evidence was reported to justify this procedure. The final,
and perhaps the most worrying aspect of the results is that Loo appears to have overfactored at the 1st order, there being nine factors identified with I 5 item salients > + 0.29;
factor V Nervous Tension consists of three high loading items, N31, N41, and N75, and
E25 (0.30), this is without any doubt a specific.
Thus it was seen as essential to carry out an analysis that took into account the points
made above, and to produce results that rested on a sounder methodological footing
(Vaughan, 1973).
e.*.,.0.
l/4--*

317

PAUL BARRETTand PAUL KLINE

318

METHOD

Subjects
235 male and 171 female undergraduates
were administered
the EPQ under group
testing conditions.
A Gallup quota sample of 600 English male adults and 598 English
female adults was also used as data for this study. The characteristics
of this sample and
details of data collection are provided in Eysenck (1979). The samples were split into six
data subsets
of total student
(N = 406) student
male (N = 171) student
female
(N = 235) total Gallup (N = 1198) Gallup male (n = 600) and Gallup female (N = 598).
E, P, N, and L scale means and standard deviations were calculated for each sex specific
sample.
Factor

Analysis

For each of the 6 raw scored data sets, phi coefficient matrices were computed.
Measures of sampling adequacy (MSA) were computed for all six data sets prior to each
level of factoring (Kaiser and Rice, 1974; Cerny and Kaiser, 1977; Meyer et al., 1977). A
principal components
analysis was carried out, using both the Kaiser-Guttman
(K-G)
and Scree test (Cattell, 1966; Cattell and Vogelmann,
1977; Horn and Engstrom,
1979) to
select those factors for retention
at the 1st order level. The factors thus retained were
rotated using the direct oblimin procedure (Jennrich and Sampson, 1966; Jennrich, 1979)
with the 6 parameter swept from -30 to 0.5 in steps of 0.5, the convergence
criterion set
at 0.00001, with a maximum
of 400 iterations
per 6. The associated overall HC (kO.1
bound) for each of these rotations was noted, the appropriate
solution being given by the
maximum HC and its associated 6. Then the rotation was again carried out around this 6
value in steps of 0.1 to fix the solution. (Direct oblimin has the attractive characteristic
that obliquity can be varied from near orthogonality
to strong obliquity by varying the
parameter 6, thus the solution is virtually unconstrained
by the rotation method.) Again
the maximum
HC determined
the solution, the minimum
overall sum of squared loadings within the hyperplanes
fixed the solution within possible equal HC plateaus. This
solution was accordingly
interpreted.
Given that a sufficient number of factors had been
extracted at this order, a 2nd order solution was undertaken.
This involved principal
component
factoring of the 1st order factor correlation
matrix, a Hendrickson-White
(H-W) grounding
operation
on the chosen unrotated
2nd order factors, then the same
direct oblimin procedures as above to yield a maximised simple structure solution. This
higher order factoring procedure would continue
as required. However, there is a real
problem with the extraction
criteria to be used at the higher order stages of factoring.
The K-G criterion, if applied to the unrotated
higher order factor eigenvalues,
will yield
results that differ markedly from the grounded
variable unrotated
higher order factor
eigenvalues
(if they can be said to represent such). The H--W operation
will obviously
boost the variance explained by the inclusion
of all item variables into that factor. In
order to resolve this apparent dilemma, it was decided that extraction
at higher orders
would be under a loose expectation
of finding representative
factors for E, N, P and L.
This involved taking out 2, 3 and 4 factors above the K-G criterion on the nongrounded
unrotated
higher order factors, even when it indicated
a lesser number.
The solution
showing the most interpretable
clarity and most closely approximating
the E, N, P and L
factorial structure was chosen. Finally, oblique rotated factor variance coefficients were
calculated for the higher order factors, and the overall significance of the simple structure
solution assessed where possible by the Bargmann
Test (1955) using the tables provided
by Kameoka and Sine (1978) given in Cattell (1978). Mean HC per solution were used in
these comparisons
as the Bargmann
test tables present values for only generalised
solutions.
RESULTS
The E,N,P and L scale means and standard
deviations
for the sex specific subject
groups are shown in Table 1, with the associated mean and standard deviation of each

Personality

factors

Table

Scale

Gallup male
N = 600

in the Eysenck

personality

1. Scale means and standard

319

questionnaire

deviations

Gallup female
N = 598

Student male
N = 171

Student female
N = 235

Mean

S.D.

Mean

S.D.

Mean

S.D.

Mean

S.D.

L
E
N
P

7.6
12.6
9.8
4.0

4.1
5.2
5.4
37

9.3
12.7
13.2
2.4

4.1
4.9
5.2
2.3

4.9
12.6
10.2
4.9

3.5
5.0
5.0
3.5

5.7
13.7
11.7
3.4

3.3
4.9
5.1
2.9

AGE

34.3

12.1

33.7

12.2

21.1

5.6

20.2

3.3

sample age. Tables 2-7 present the first order pattern solution and associated statistics
for each of the six data sets; only salient loadings > f 0.3 appear in the tables. Factor
names were chosen on the basis of reasonably clear loading patterns. Those factors with
apparently complex patterns were left unnamed. Premature naming of such factors is not
helpful, rather, the factors are presented purely for future reference in any other item
factoring of the EPQ. (The stability of such unexpected factors was not known at the
time of writing.) The rotations for each final 6 all reached convergence at less than the
400 iterations specified. All solutions were shown to have reached a satisfactory simple
structure using Bargmanns test (p < 0.01). All 1st order factor correlations were less than
f 0.26.
All solutions were subsequently taken to the 2nd order, using the procedures described
above. Tables 8-13 present the six 2nd order pattern solutions. Once again all rotations
for the retained 6 reached convergence at less than 400 iterations. Factor naming was, for
these factors, straightforward. Pearson Y and Tucker congruences were subsequently
computed between all factors in all six solutions in order to aid factor matching across
data sets. Coefficients >0.75 were taken as demonstrating sufficient similarity for two
factors to be said to be jointly represented. Factors 3 and 4 in the student female, factor 4
Table 2. Total student
Extraversion
Factor 1
El
E5
El0
El4
El7
E21
E25
E29
E32
E40
E42
E45
E49
E52
E60
E64
E70
E82
E86
Eigenvalue

0.303
0.636
0.664
0.516
0.549
0.686
0.533
0.589
0.540
0.498
0.621
0.486
0.387
0.642
0.325
0.307
0.533
0.584
0.649

Moody irritability
Factor 2
N3
N7
N15
N23
N3l
N34
N41
N54
N58
N62
N68
N75
N77
N84

(unrotated)
7.686

Total solution HC = 275.

Total solution HC% = 51.
Mean HC = 46.
6 = 4.1.
MSA = 0.75.

0.558
0.614
0.412
0.655
0.337
0.386
0.364
0.381
0.584
0.434
0.426
0.380
0.409
0.511

5.694

Social desirability
Factor 3
J_4
L8
L16L24
L28
L39
L44
L48
L59
L63
N66
L78
L89

0.545
0.450
0.499
0.369
0.307
0.429
0.407
0.347
0.553
0.511
-0.346
0.485
0.490

4.210

primaries
Anxious worrying
Factor 4
Pll
N12
N19
N27
N3l
P33
N34
N38
N41
P53
P71
N72
P79
N80
P83

-0.439
0.483
0.380
0.449
0.339
-0.437
0.413
0.430
0.315
- 0.462
-0.327
0.542
-0.331
0.465
-0.339

3.188

Caution
Factor 5
P2
P6
P9
L13
P22
P37
P57
P67
P74
L81

0.316
0.417
0.517
-0.320
0.381
0.475
0.529
0.436
0.644
- 0.498

2.329

Factor
L20
P26
P30
N41
E56
P65
P76
L85
P87

6
0.324
0.375
0.552
0.351
0.335
0.447
0.385
0.344
0.430

2.111

N3
N7
N23
P30
N38
N41
N54
N58
N62
P65
P76
N77
P87

(unrotated)
8.225

0.551
0.564
0.327
0.505
0.447
0.415
0.552
0.393
0.709
0.441
0.367
0.712
0.650

6.213

0.460
0.507
0.371
0.548
0.307
0.403
0.465
0.380
0.412
0.573
0.555
0.51 I
0.483

Moody irritability
Factor 2

Total solution- HC = 418

Total solution HC,, = 52.
Mean HC = 46.
6 = - 1.2.
MSA = 0.54.

Eigenvalue

E5
El0
El7
E21
E29
E32
E40
E42
E45
E49
E52
E70
E86

Sociability
Factor I

L4
L8
L16
L24
L28
L39
L44
L59
L63
L78
L89

4.076

0.571
0.547
0.462
0.422
0.476
0.577
0.559
0.563
0.485
0.572
0.464

Social desirability
Factor 3
P6
N12
N19
N27
N34
N38
N47
P53
N66
N72
NXO

3.557

-0.379
0.673
0.463
0.465
0.492
0.406
0.343
-0.513
0.324
0.674
0.458

Anxious worrying
Factor 4

2.805

0.315
0.575
0.352
0.341
0.646
0.367
0.460
-0.480
0.523
0.516
-0.399

Liveliness
Factor 5

male primaries

El
El4
E21
P22
E25
E42
ES2
L55
E60
E82
N88

Table 3. Student

PI1
N15
P22
P26
P33
P53
N68
P71
P79
P83

2.552

0.440
0.476
0.356
0.55 1
0.607
0.350
0.331
0.492
0.331
0.496

Insensitivity
Factor 6
P9
L13
El7
L20
P22
N47
E56
P57
N58
P74
LX1

2.340

0.412
-0.350
0.376
-0.304
0.309
0.330
- 0.459
0.439
0.349
0.604
-0.541

Factor

P18
E36
P37
P43
P50
P61
P67
P76

8
0.404
0.356
0.349
0.379
0.328
0.503
0.668
0.311

Factor

L20
N31
L35
L48
L51
L55
N75
N84
L85

-0.412
-0.312
0.527

0.301
0.356
0.533
0.359
0.546
0.368

Factor

Y
2
1
3

9
:

(unrotated)
7.889

0.319
0.595
0.686
0.453
0.620
0.683
0.561
0.648
0.525
0.302
0.478
0.631
0.482
0.321
0.645
0.374
0.378
0.505
0.538
0.734

5.986
3.932

0.434
0.495
0.455
0.32 1
0.458
0.302
0.373
0.529
0.544
0.416
0.300
0.367
- 0.305
0.307

L4
LB
L24
L28
L35
L39
L48
L51
L59
L63
L37
L78
N80
L89

N12
N15
Nl9
N27
P30
N31
N34
E36
N38
N41
N47
P53
N72
N75
N77
N80

0.447
0.343
0.498
0.545
0.333
0.626
0.648
-0.325
0.566
0.631
0.320
-0.410
0.571
0.672
0.477
0.454

Social desirability
Factor 3

Anxious worrying
Factor 2

Total solution HC = 440.

Total solution HC% = 54.
Mean HC = 49.
6 = 0.0.
MSA = 0.66.

Eigenvalue

El
E5
El0
El4
El7
E21
E25
E29
E32
E36
E40
E42
E45
E49
E52
E56
E64
E70
ES2
E86

Extraversion
Factor 1
N3
N7
N15
N23
N58
N62
P71
N84

3.354

0.585
0.665
0.321
0.586
0.595
0.383
0.306
0.606

P2
P6
P9
L13
P50
P53
P57
P74
L81

2.653

0.333
0.427
0.375
-0.401
0.312
0.357
0.662
0.682
-0.355

Caution
Factor 5

female primaries

Moody irritability
Factor 4

Table 4. Student

P2
L8
P18
L20
P37
E45
P67
E70
P87

Factor

2.347

0.373
0.339
0.671
0.452
0.322
0.366
0.669
0.348
0.332

6
P22
P26
P46
P65
P76
P83
L85

Factor

2.148

0.352
0.629
0.381
0.338
0.527
0.583
0.420

7
Pll
N19
P30
P37
L55
N68
NXX
L89
P90

2.052

0.321
-0.301
0.372
0.377
-0.305
0.393
-0.337
0.320
0.405

Factor

-0.376
0.311
-0.301
0.465

E52
E56
P79

9
El

Factor

5
z
Y5

(unrotated)
7.690

Eigenvalue

P6
N12
N19
N27
N31
N34
N38
N41
N47
P53
N66
N72
N75
N80

6.235

-0.374
0.608
0.553
0.384
0.420
0.569
0.510
0.367
0.321
- 0.402
0.406
0.652
0.383
0.645

Anxious worrying
Factor 2

Total solution HC = 492.

Total solution HC% = 61
Mean HC = 55.
6 = -0.8.
MSA = 0.87.

0.334
0.420
0.483
0.665
0.380
0.584
0.678
0.511
0.403
0.445
0.326
0.377
0.340
0.720
0.385
0.529
0.368

E5
El0
El4
El7
E21
E25
E29
E32
E36
E!40
E42
E45
E49
E52
E70
ES2
ES6

Extraversion
Factor 1

L4
L8
L16
L20
L24
L28
L39
L44
L48
L51
L59
L63
L73
L81

4.603

0.336
0.613
0.400
0.326
0.577
0.377
0.442
0.525
0.500
0.313
0.592
0.610
0.332
0.301

Social desirability
Factor 3
N3
N7
N15
N23
N58
N62
N68
N77
N84

3.800

0.665
0.571
0.370
0.676
0.544
0.524
0.307
0.336
0.588

Moody irritability
Factor 4
P18
P30
P43
P65
P67
P76
N77
P87

1.975

1.757

-0.321
0.514
0.429
0.547
0.322
0.423
0.442
0.346
0.572

Insensitivity
Factor 6
L4
P11
P26
P33
P46
P71
P79
P83
P90

primaries

0.332
0.500
0.364
0.546
0.415
0.48 1
0.313
0.542

Factor

Table 5. Total Gallup

El
E5
El0
E40
E42
E45
E60
E64
E70
L85
ES6

1.677

0.400
0.306
0.408
0.322
0.369
0.394
0.443
0.384
0.382
0.357
0.415

Social liveliness
Factor 7
P9
Pt8
P22
P43
P57
P67
P74

1.613

0.308
0.538
0.356
0.308
0.609
0.498
0.626

Caution
Factor 8

0.378
0.430
0.404

N4l
E42
N75

1.461

0.300
0.459
0.320
0.347

9
P2
ES
E2l
N3l

Factor

Y
E

P
5

P6
N12
N19
N27
N31
N34
N38
N41
N47
P53
N66
N72
N75
N80
N88

(unrotated)
7.902

0.494
0.542
0.509
0.684
0.529
0.625
0.659
0.537
0.375
0.609
0.517
0.520
0.330
0.729
0.576
0.483
0.534

6.803

-0.305
0.588
0.540
0.436
0.584
0.617
0.498
0.524
0.365
-0.347
0.304
0.645
0.541
0.612
0.301

Anxious worrying
Factor 2

Total solution HC = 469.

Total solution HC:d = 58.
Mean HC = 52.
6 = -0.2.
MSA = 0.83.

Eigenvalue

E5
El0
El4
El7
E21
E25
E29
E32
E36
E40
E42
E45
E49
E52
E70
E82
E86

Extraversion
Factor 1

N3
N7
N15
N23
N58
N62
N68
N77
N84

4.429

0.652
0.608
0.372
0.707
0.525
0.545
0.360
0.315
0.589

Moody irritability
Factor 3
E5
Pll
P26
P33
E42
P61
P71
L85
P90

4.215

0.375
0.566
0.509
0.405
0.384
0.441
0.350
0.316
0.554

Factor

2.240

0.371
0.425
0.333
0.493
0.395
0.499
0.516
0.378
-0.445

Interpersonal
social
desirability
Factor 5

L4
L8
L16
L24
L39
L59
L63
L69
P79

Table 6. Male Gallup

P30
P43
P50
P53
N62
P65
P76
N77
P87

1.915

0.428
0.384
0.329
0.325
0.303
0.578
0.505
0.335
0.563

Psychotic depression
Factor 6

primaries

L8
L13
L16
L20
L28
L44
L78
L81
L85
L89

1.771

0.386
0.334
0.300
0.550
0.454
0.348
0.582
0.339
0.447
0.307

Abstract morality
Factor 7
P9
P18
P22
P37
P43
P57
P67
P74

1.609

0.434
0.570
0.361
0.316
0.386
0.634
0.553
0.626

Caution
Factor 8

L35
L48
L51
L73

Factor

0.350
0.379
0.447
0.366

!?
0
7
6

2
;I
2

(unrotated)
7.364

0.458
0.493
0.546
0.655
0.490
0.523
0.639
0.481
0.432
0.499
0.42 1
0.451
0.425
0.701
0.498
0.562
0.402

5.963

0.365
0.334
0.385
0.430
0.350
0.685
-0.364
0.424
0.375
0.697
0.328
0.357
0.382
0.419
0.357
0.745
0.388
0.425
0.377

Neuroticism
Factor 2

N3
N7
N15
N23
N27
N31
E32
N34
N38
N41
N47
N54
N58
N62
N68
N75
P76
N77
P87

Total solution HC = 366.

Total solution HC% = 58.
Mean HC = 52.
6 = -0.4.
MSA = 0.80.

Eigenvalue

E5
El0
El4
El7
E21
E25
E29
E32
E36
I?,40
E42
E45
E49
E52
E70
ES2
E86

Extraversion
Factor 1
L4
L8
L13
L16
L20
L24
L28
L35
L39
L44
L48
L51
L55
L59
L63
L73
L78
L89

3.863

0.363
0.615
0.343
0.500
0.368
0.530
0.368
0.371
0.450
0.485
0.507
0.346
0.307
0.505
0.596
0.348
0.346
0.342

Social desirability
Factor 3

Gallup

P6
N12
N19
N34
N38
P53
N66
N72
N80
N88

2.912

0.546
0.542
0.434
0.378
- .0.413
0.439
0.495
0.580
0.328

_ -0.315

Anxious worrying
Factor 4

Table 7. Female

primaries

P18
P30
P33
P43
P65
P67
P71
P87

Factor

1.924

0.506
0.361
0.426
0.48 1
0.434
0.583
0.314
0.367

5
El
El0
E42
E60
E64
L85
E86

1.864

0.482
0.384
0.301
0.496
0.412
0.377
0.432

Social liveliness
Factor 6
N3
N7
N23
P57
N58
P74
L81
N84

0.395
0.392
0.342
0.479
0.323

1.742

0.588
-0.321
0.323

Factor

P
F

coeff.

El
E5
El0
El4
El7
E21
E25
E29
E32
E36
E40
E42
F.45
E49
E52
E56
E60
E64
E70
ES2
ES6

HC = 166.
HC% = 46.

factor variance

Total solution
Total solution
6 = -25.2.
MSA = 0.51.

Rotated

Extraversion
Factor 1

Var. title

6.346

0.310
0.646
0.653
0.400
0.480
0.659
0.455
0.513
0.519
0.324
0.533
0.612
0.550
0.416
0.539
0.404
0.364
0.311
0.608
0.540
0.674

Loading

Neuroticism
Factor 2
N3
N7
N12
N15
N19
N23
N27
P30
N31
N34
N38
N41
N47
N54
N58
N62
P65
N66
N68
N72
N75
P76
N77
N80
PSO
N88

Var. title

5.928

0.403
0.357
0.375
0.515
0.438
0.456
0.453
0.444
0.554
0.598
0.533
0.635
0.345
0.395
0.359
0.358
0.406
0.321
0.384
0.433
0.556
0.392
0.543
0.382
0.362
0.367

Loading

Table 8. Total student

Psychoticism
Factor 3

secondaries

P2
P6
P9
Pll
N12
P18
P22
P26
P33
E36
P37
P46
P53
P67
P71
N72
P74
N80

Var. title

4.362

0.389
0.546
0.457
0.300
-0.304
0.367
0.444
0.308
0.382
0.365
0.486
0.378
0.364
0.422
0.351
-0.384
0.330
- 0.356

Loading
Social
desirability
Factor 4

0.353
0.388
0.357
0.390
0.334
0.348
0.372
L13
L16
L20
L24
L35
L39
L44
L48
L51
L59
L63
L78
L81
L85
L89

3.926

0.439

0.348

0.410
0.370

0.399
0.369
0.426
0.409

0.404
0.477

Loading

L4
L8

Var. title

c
w
2
K
0
8
g
E
=:
G

$
I:
G
3
g
2
5

3
;I

coeff.

E5
El0
El4
E21
E25
E29
N31
E32
P33
E36
E40
E42
E45
E49
E52
E56
E60
E70
E82
E86

HC = 153.
HCU;;, = 43

factor variance

Total solution
Total solution
6 = -6.3.
MSA = 0.57.

Rotated

Extraversion
Factor 1

Var. title

6.463

0.644
0.669
0.435
0.662
0.429
0.388
-0.412
0.514
0.309
0.304
0.425
0.640
0.511
0.526
0.530
0.333
0.461
0.543
0.582
0.569

Loading

Neuroticism
Factor 2
N3
N12
N15
N19
N23
P26
N27
P30
N31
N34
N38
N41
N54
N62
P65
N66
N68
N72
N75
P76
N17
N80
P83
P87
N88

Var. title

5.921

0.334
0.425
0.507
0.392
0.458
0.315
0.490
0.477
0.411
0.567
0.419
0.504
0.358
0.412
0.519
0.379
0.394
0.372
0.383
0.449
0.547
0.331
0.329
0.309
0.385

Loading

Table 9. Student

.,I

Social
desirability
Factor 3

male secondaries

L4
L13
L16
t20
L39
L44
L48
P50
L51
N58
L59
L63
PI9
L81
L85
L89

Var. title

4.65 1

0.509
0.374
0.400
0.355
0.400
0.507
0.518
0.352
0.309
-0.386
0.424
0.429
-0.374
0.507
0.477
0.424

Loading

Factor

P2
P6
P9
Pll
P18
P22
E36
P37
P46
P53
P57
P61
P67
N72
P74

Var. title

0.411
0.573
_ .0.364

4.450

0.382

Y
z

0.343
0.399

7
rc

6
a

;P
5

0.401
0.518
0.523
0.331
0.352
0.421
0.439
0.446
0.355

Loading

coeff.

El
E5
El0
El4
El7
L20
E21
E?5
P26
E29
P30
E32
P33
E36
E40
E42
E45
P46
E49
ES6
E60
E70
E82
P83
E86
P87

HC = 144.
HCY; = 40.

factor variance

Total solution
Total solution
6 = 0.6.
MSA = 0.55.

Rotated

Extraversion
Factor 1

Var. title

6.006

0.324
0.530
0.467
0.328
0.361
0.411
0.489
0.371
0.466
0.402
0.401
0.336
0.322
0.330
0.448
0.455
0.461
0.330
0.333
0.514
0.382
0.572
0.369
0.395
0.532
0.397

Loading
Neuroticism
Factor 2
N3
N7
N15
N23
P30
N31
N34
N38
N41
N47
N54
N58
N62
P65
N68
N72
N75
P76
N77
N84
N88

Var. title

5.555

0.435
0.454
0.480
0.544
0.349
0.546
0.527
0.332
0.569
0.383
0.464
0.439
0.473
0.356
0.413
0.311
0.601
0.324
0.511
0.452
0.313

Loading

Table 10. Student

Factor

female secondaries

P3
P9
L13
L24
L28
E29
P37
L39
P53
L55
P57
L59
P61
P71
PI4
L81
L85

Var. title

4.322

-0.339
- 0.460
0.529
0.370
0.354
-0.455
-0.463
0.333
-0.387
0.353
-0.365
0.402
-0.334
-0.318
-0.456
0.494
0.346

Loading

Factor

L4
L8
Pll
N12
L16
N19
E29
L48
N66
L73
L78
N80
N88
L89

Var. title

3.831

-0.418
0.460

-0.366
0.346
0.440
-0.511

0.426
- 0.423
- 0.493
0.351

0.302
0.376
0.381
-0.355

Loading

c
w
z
A

z
5

2
F

2
=
G

I?
a

0.414

0.412
0.469

0.483
0.358

0.434
0.485
0.637

0.548
0.448
0.375
0.431

N47

N54
N58

N62
N66

N68
N72
N75

N77
NSO
N84
N88

6.561

0.632
0.578
0.601

N34
N38
N41

coeff.

0.524
0.552
0.425
0.492
0.482
0.615
0.520
0.607

Loading

N3
N7
N12
N15
N19
N23
N27
N31

HC = 181.
HC% = 50.

factor variance

Total solution
Total solution
6 = 0.2.
MSA = 0.54.

Rotated

Neuroticism
Factor 1

Var. title

Extraversion
Factor 2

E60
E82
E86

E52
E56

E45
E49

E42

E32
E36
E40

El
E5
El0
El4
El7
E21
E25
E29

Var. title

5.937

0.392
0.504
0.634

0.549
0.375

0.618
0.340

0.558

0.417
0.391
0.613

0.324
0.552
0.668
0.479
0.504
0.557
0.449
0.456

Loading

Table 11. Total Gallup

Psychoticism
Factor 3

secondaries

P83
P87
P90

P67
P71
P76

P61
P65

P53
P57

P50

P37
P43
P46

P6
P9
Pll
Pl8
P22
P26
P30
P33

Var. title

4.849

0.406
0.310
0.434

0.374
0.436
0.390

0.345
0.426

0.398
0.339

0.307

0.448
0.485
0.302

0.373
0.323
0.488
0.404
0.431
0.550
0.383
0.515

Loading
Social
desirability
Factor 4

L89

L78
L81
L85

L69
L73

L59
L63

L51

L39
L44
L48

L4
L8
L13
L16
L20
L24
L28
L35

Var. title

4.355

0.345

0.454
0.438
0.384

0.387
0.394

0.387
0.490

0.383

0.355
0.390
0.442

0.336
0.525
0.403
0.390
0.435
0.465
0.337
0.418

Loading

T:
s
m

;
CL
w

?
$

7
5

coeff.

N3
N7
N12
N15
N19
N23
N27
N31
N34
N38
N41
N47
N54
N58
N62
N68
N72
N75
N77
N80
N84
N88

HC = 168.
HC% = 47.

factor variance

Total solution
Total solution
6 = 0.5.
MSA = 0.56.

Rotated

Neuroticism
Factor 1

Var. title

6.373

0.557
0.550
0.403
0.506
0.482
0.664
0.517
0.566
0.625
0.571
0.518
0.452
0.440
0.472
0.506
0.410
0.372
0.542
0.548
0.435
0.410
0.486

Loading

Extraversion
Factor 2
El
E5
El0
El4
El7
E21
E25
E29
P30
E32
E36
E40
E42
E45
E.49
E52
E56
E60
E70
E82
E86

Var. title

Table

5.919

0.353
0.522
0.678
0.446
0.341
0.500
0.371
0.339
0.326
0.365
0.438
0.601
0.528
0.684
0.322
0.421
0.388
0.411
0.694
0.489
0.651

Loading

12. Gallup

Psychoticism
Factor 3

male secondaries

P2
L4
P6
P9
Pll
P18
P22
P26
P30
P33
P37
P43
P46
P50
P53
P57
P61
P65
P67
P71
P74
P76
P83

Var. title

5.546

0.340
_ 0.309
0.384
0.415
0.543
0.422
0.482
0.650
0.340
0.535
0.497
0.513
0.313
0.353
0.403
0.430
0.415
0.402
0.336
0.421
0.378
0.383
0.419
0.463

Loading
Social
desirability
Factor 4
L4
L8
L13
L16
L20
L24
L28
L35
L39
L44
L48
L55
L59
L63
L69
L73
L78
L81
L85
L89

Var. title

0.373
0.410
0.340
0.498
0.524

%
=f
G
?
D
5.
2
$.
;p

0.424
0.496
0.49 1
0.351

4.609

B
7

0.390
0.333

c
Yz
R
k?

%
5

2
z
:
!%
G
2
I
0.315
0.366

0.461
0.487
0.381

0.401
0.553
0.432
0.424

Loading

coeff.

N3
N7
N12
N15
N19
N23
N27
N3l
N34
N38
N41
N47
P53
N54
N58
N62
N66
N72
N75
N77
N80
N84
N88

HC = 180.
HC% = 50.

factor variance

Total solution
Total solution
6 = -21.0.
MSA = 0.52.

Rotated

Neuroticism
Factor 1

Var. title

5.848

0.504
0.428
0.461
0.437
0.505
0.522
0.489
0.464
0.622
0.549
0.487
0.459
-0.333
0.315
0.378
0.407
0.441
0.558
0.538
0.485
0.507
0.337
0.420

Loading

Extraversion
factor J
El
E5
El0
El4
El7
E21
E25
E32
E40
E42
E45
E52
E56
E60
E64
E70
E82
E86

Var. title

Table

5.286

0.524
0.646
0.371
0.433
0.524
0.354
0.349
0.542
0.558
0.511
0.432
0.388
0.513
0.403
0.514
0.426
0.614

Social
desirability
Factor 3

female secondaries

0.440

Loading

13. Gallup

N7
L8
L13
L16
L20
L24
L35
L44
L48
L51
P57
L63
L69
Ll3
P74
L78
L8l
L85
L89

Var. title

3.989

-0.333
0.473
0.405
0.302
0.353
0.41 I
0.426
0.331
0.402
0.420
-0.351
0.367
0.348
0.376
- 0.424
0.401
0.418
0.331
0.335

Loading

Factor

P2
P18
P26
P30
P33
P43
E45
P53
P65
P67
N68
P71
P76
P79
P87

Var. title

3.664

0.323
0.348
0.310
0.401
0.406
0.402
0.309
0.388
0.454
0.446
0.342
0.375
0.419
0.308
0.492

Loading

Personality

Table

factors

14. Numbers

in the Eysenck

personality

of E, N, P and L items
second order factors
E

Total students
Male students
Female students

211
181
19,

232
192
182

Total Gallup
Male Gallup
Female Gallup

2%
2%
182

23,
22,
22,

Sample

Where
No. of
No. of
No. of
No. of

subscript
E scale
N scale
L scale
P scale

denotes factor
items = 21.
items = 23.
items = 21.
items = 25.

extraction

questionnaire

loading

L
17,
133
L/P
ITEMS
2%
204
1%

331

on the

P
143
134
N/L
1TEMS
223
23,
134

order.

in the Gallup female, and factor 4 in the student male, could not be matched across all
six studies. No tests of simple structure were made on these data as there are no tabled
probability
values available for this size of solution. Table 14 presents a summary of the
salient scale items of E, N, P and L, appearing in the 2nd order factor solutions. All 2nd
order factor correlations
were less than +0.2 excepting that between factors 1 and 3 in
the female student data. No further analysis took place.
DISCUSSION

Before discussing these results, it is perhaps of value to state the procedures

adopted
for reducing the 49 total 1st order factors extracted to a common and specific subset.
Those factors on which the same loading pattern was clearly discernable
across all six
studies were considered
replicable
primaries.
The factors remaining
which could be
named were considered
specific, just how specific depended upon cross matches by sex
and by status (adult-student).
For the 24 total 2nd order factors, this procedure was not
required as the patterns were all so similar in but a few cases noted above. In all six 1st
order studies, two factors were common, Moody Irritability
and Anxious Worrying;
two
factors were common to five of these studies, Extraversion
and Social Desirability;
in the
student male sample, Sociability was a more accurate interpretation
of the item loadings
than was the global concept of Extraversion;
in the Gallup male study, Social Desirability appeared to split into two meaningful factors of Interpersonal
Social Desirability
and
Abstract
Morality.
Notably
for the Extraversion
factor, practically
the entire scale
emerged at the 1st order, supporting
in part Eysenck and Eysencks (1968) claim that E,
N and P can be found at the 1st as well as the 2nd and 3rd order. The remaining
1st
order factors that were named did not appear to be status specific or sex specific, rather
they were virtually study specific.
The 2nd order factors from the six data sets are in one sense quite spectacular.
From
Table 14, it can be seen that practically
all the items of the E, N and L scales were
recovered across the data sets. The P scale items, however, appeared with such clarity
only in the total Gallup and male Gallup samples. Factor similarity analysis demonstrating that the total Gallup and total student factors were very similar in content if not in
quantity of salients.
Where the factors (subsequently
named using Eysencks descriptions)
did not appear,
two interpretations
are available: either these factors are approximating
the concepts or
they are sample specific item patterns.
Noting
the samples in which no matching
occurred, factors 3 and 4 in the student female sample appear to represent two sets of
covarying items, L and P, L and N. Why this should be so is not at all clear [Loo (1979)
reported such evidence of covariance
between the L and P scales in his student female
subsample].
However, the difficulty of obtaining
a clear P factor is probably due to the
sensitivity of the factor to the item score patterns, notably the mean of the raw scores on
the P scale for both female samples (student and Gallup) is very low for a 25 item

PAUL BARRETT and PAUL KLINE

332

scale-the
score distribution
is greatly distorted;
there is simply insufficient variance to
be factored. The missing P factor in the student male sample may perhaps be attributed
to the very low variable to subject ratio in this group (1: 1.9).
Interestingly,
a factor of Impulsivity
did not appear in any of the data sets (unlike
Loos results). However, this is not very surprising as Eysenck and Eysenck (1977, 1978)
stated that general Impulsivity
items were specifically excluded from the EPQ E scale.
They did this because of the correlation
of Impulsivity
with both E and P, which would
yield an artifactual
correlation
between factors E and P in a factor analysis. They also
concluded that general Impulsivity
was related more to P and abnormality
rather than to
Extraversion
per se; additionally,
the Impulsivity
concept was split into four separate
components
and thus viewed as a factor complex rather than a unitary dimension,
contrary to Guilfords (1975, 1977) statements
and primary factor constructs. Thus Loos
factor of Impulsivity can be seen as the result of premature naming of a specific. Factors
with only four or less salients > f 0.3 as a 90 item questionnaire
cannot be regarded as
common factors (Nunnally,
1978).
On a final methodological
note, there is one possible difficulty in acceptance
of these
results, as they stand. The charge of underfactoring
may be levelled against these analyses, essentially at the 1st order extraction
phase. The arguments
against this have been
put forward above-the
emergence of small factors loading on a few items (specifics) and
the use of the scree test for determining
the number of factors to be extracted within the
K-G bound. Looking at Tables 2 to 7, it is apparent that the last one or two factors are
becoming in some cases item cluster specifics, and in others, producing
dissociated item
factors which cannot be interpreted.
Perhaps the strongest argument is that at the 2nd
order, extremely clear results were demonstrated
across all 6 data sets, results as would
generally be expected by Eysenck and Eysencks original studies on these factors.
Also of a purely methodological
interest, is the factorial behaviour
associated
with
MSA coefficients. An MSA coefficient >0.65 indicates that the data are relatively suitable for factoring, the absolute lower limit bemg 0.5. Generally,
for the 1st order solutions, the clarity of solution
remained
effectively the same, given the range of MSA
coefficients. For the 2nd order solutions,
all MSAs were very low, yet clear replicable
structures resulted. However, overall the MSA seems to indicate the degree of structural
clarity, rather than any severe disruption
or distortions.
In conclusion
the study may be summarised
as containing
major results:
(i) Eysencks factors of E, N, P and L have been found to exist at the 2nd order level
of factoring. The P factor was not found satisfactorily
in the female samples, or in
the lowest variable to subject ratio sample of male students.
(ii) E appears the strongest factor at the 1st order, changing little if any at the 2nd
order. N was the strongest factor only at the 2nd order in the Gallup sample
subsets.
(ii) Two 1st order factors were found to exist across all 6 data sets, Moody Irritability
and Anxious Worry. Two others appeared in 5 of the 6 data sets, Extraversion
and Social Desirability.
(iv) Variable to subject ratios (ranging from 1: .9 to 1: 13) did not grossly affect the
resultant 2nd order solution, rather a firming up of loading pattern was observed.
Acknowledgement-Our

thanks

are extended

to Professor

H. J. Eysenck

for kindly providing

the Gallup

sample

data used in this study.

REFERENCES
BARGMANN R. (1954) Signijikantz-Vnrersuchungen
der.Einfachen
Strukter in der Facroren Analyse. Mitteilungsblatt fur Mathematische
Statistik. Physica Verlag, Wurzburg.
BUROS 0. K. (editor) (1978) The Eighth Mental Measurements
Yearbook. Gryphon
Press, 1978.
CATTELL R. B. (1966) The Scree test for the number of factors. Multivariate
Behau. Res. 1, 14@161.
CATTELLR. B. (1978) The Scientific Use of Factor Analysis in Behavioural
and Life Sciences. Plenum Press,
New York.

Personality

factors

m the Eysenck

CATTELLR. B. and V~GELMANNS. (1977) A comprehensive

personality

questionnaire

333

trial of the scree and KG criteria for determining

the
number of factors. Multivariate
Behav. Rex 12, 289-325.
CATTELL R. B. and WHITE P. 0. (1962) The use of higher order personality
factors in relation to variables.
Unpublished
manuscript.
CERNY B. A. and KAISER H. F. (1977) A study of a measure of sampling adequacy
for factor-analytic
correlation
matrices. Multivariate Behav. Res. 12,43-47.
EYSENCK H. J. (1977) Personality
and factor analysis: a reply to Guilford. Psychol. Bull. 84, 405411.
EYSENCK H. J. (1978) Superfactors,
P, E, and N in a comprehensive
factor space. Multivariate Behav. Res., 13,
475481.
EYSENCK H. J. (1979) Personality
factors in a random sample of the population.
Psychol. Rep. 44, 1023-1027.
EY~ENCK H. J. and EYSENCK S. B. G. (1968) A factorial study of psychoticism-as
a dimension of personality.
Multivariate
Behav. Res. Special Issue, pp. 15-31.
EY~ENCK H. J. and EY~ENCK S. B. G. (1975) Manual of the Eysenck Personality Questionnaire. Hodder &
Stoughton,
London.
EYSENCK H. J. and EYSENCKS. B. G. (1976) Psychoticism as a Dimension of Personality. Hodder & Stoughton,
London.
EYSENCK S. B. G. and EYSENCKH. J. (1977) The place of impulsiveness
in a dimensional
system of personality
description.
Br. J. sot. c/in. Psycho\. 16, 57-68.
EYSENCK S. B. G. and EYSENCK H. J. (1978) Impulsiveness
and Venturesomeness:
their position in a dimensional system of personality
description:
P&ho/.
Rep. 43, 1247-1255.
GUILFORD J. P. (1975) Factors and factors of personality.
Psycho/. Bull. 83, 802-814.
GUIDFORD J. P.(197$ Will the real factor oiextraversion-introversion
please stand up? A reply to Eysenck.
Psychol. Bull. 84, 412416.
HENDRICKSON A. E. and WHITE P. 0. (1966) A method for the rotation of higher order factors. Br. J. math. stat.
Psychol. 19,97-103.
HORN J. L. and ENCSTROM R. (1979) Cattells scree test in relation to Bartletts chi square test and other
observations
on the number of factors problem. Multivariate Behav. Res. 14, 283-300.
JENNRICH R. I. (1979) Admissible values of y in direct oblimin rotation. Psychometrika 44, 173-177.
JENNRICH R. 1. and SAMP~QN P. F. (1966) Rotation for simple loadings. Psychometrika, 31, 313-323.
KAISER H. F. and RICE J. (1974) Little Jiffy, mark IV. Educ. psycho/. Meas. 34, 111-117.
KAMEOKA V. and SINE L. (1978) An extension of Bargmanns
tables and a computer
program
for testing the
significance of simple structure in factor analysis. Manuscript
submitted for publication.
Loo R. (1979) A psychometric
investigation
of the Eysenck Personality
Questionnaire.
J. Person. Assess. 43,
54-58.
MEYER E. P., KAISER H. F., CERNY B. A. and GREEN B. F. (1977) MSA for a special Spearman
matrix.
Psychometrika 42, 153-156.
NUNNALLY J. C. (1978) Psychometric Theory (2nd edn). McGraw-Hill,
New York.
VAUGHAN D. S. (1973) The relative methodological
soundness of several major personality
factor analyses. J.
Behav. Sci. 1, 305-313.